Treatment delivery was centrally reviewed by the study chair and medical physics co-chair. Success was defined as having a review of "per protocol" or "minor variation". A 70% success rate was required for continuation to the phase III component.

• Target volume compliance in is defined as the following levels of target coverage in reference to dose volume:

  • Per protocol: ≥ 90%
  • Minor variation: 80%-90%
  • Major deviation: <80% of of dose volume.

• Image-Guided Radiotherapy (IGRT) compliance is defined as the following differences in IGRT images between simulation/planning and treatment, as well as at the end of treatment:

  • Per protocol: <2 mm
  • Minor variation: 2mm - 3mm
  • Major deviation: > 3mm

Patients enrolled in the phase II component who were eligible and had treatment review information.

Pain is measured by the Numerical Rating Pain Scale (NRPS) a numerical 11-point scale (0-10) with 0 = no pain, 1-4 = mild, 5-6 = moderate, and 7-10 = severe with 10 the worst pain imaginable. Pain response is calculated as 3 month score - baseline score with a positive value indicating increased pain and a negative value indicating decreased pain. Patients with complete or partial pain response as defined below are considered responders. The index site is the lesion with the highest baseline (day of radiosurgery) pain score. If multiple sites have the same baseline pain score, the index site is the most cephalad lesion.

• Complete response: post-treatment pain score of 0 at the index site, no increase in narcotic pain medication, and no increase in pain score at the secondary treated site(s).
• Partial response: post-treatment improvement of at least 3 points at the index site, no increase in narcotic pain medication, and no increase in pain score at the secondary treated site(s).

Eligible participants randomized in the phase III component with both baseline and 3 month NRPS scores

Pain is measured by the NRPS, a numerical 11-point scale from 0=no pain to 10=worst pain imaginable. Magnitude of pain response = post-baseline score - baseline score with a positive value indicating increased pain and a negative value indicating decreased pain. Pain response begins when a patient improves at least 3 points at the index site. (Although pain medication and pain response were part of the protocol definition of pain response, they were unable to be assessed accurately during weeks 1-3 and therefore were not included in the final definition used here.) Pain response time is defined as time from randomization to the date of pain response, last known follow-up (censored), or death (censored). Rates of participants with pain response are estimated using the Kaplan-Meier method. The distributions of pain response times are compared between the arms. Two-year rates are provided.

Eligible participants randomized in the phase III component with pain response data

Pain is measured by the NRPS, a numerical 11-point scale from 0=no pain to 10=worst pain imaginable. Amount of pain response = post-baseline score - baseline score with a positive value indicating increased pain and a negative value indicating decreased pain. Pain response begins when a patient improves at least 3 points at the index site. Pain response ends when when the pain score increases by 3 points (relapse). (Although pain medication and pain response were part of the protocol definition of pain response, they were unable to be assessed accurately during weeks 1-3 and therefore were not included in the final definition used here.) Pain relapse time is defined as time from randomization to the date of first failure, last known follow-up (censored), or death (censored). Rates of participants with pain relapse are estimated using the Kaplan-Meier method. The distributions of pain relapse times are compared between the arms. Two-year rates are provided.

Eligible participants randomized in the phase III component who experienced partial or complete pain response

Common Terminology Criteria for Adverse Events (version 3.0) grades adverse event (AE) severity from 1=mild to 5=death. Summary data is provided in this outcome measure.; See Adverse Events Module for specific adverse event data.

Eligible participants randomized in the phase III component with adverse event information

Failure is defined as a vertebral compression fracture. Failure time is defined as time from randomization to the date of first failure, last known follow-up (censored), or death without compression fracture (competing risk). Failure rates are estimated using the cumulative incidence method. The distributions of failure times are compared between the arms. Two-year rates are provided.

Eligible participants randomized in the phase III component

Failure is defined as spinal cord damage. Failure time is defined as time from randomization to the date of first failure, last known follow-up (censored), or death without compression fracture (competing risk). Failure rates are estimated using the cumulative incidence method. The distributions of failure times are compared between the arms. Two-year rates are provided.

Eligible participants randomized in the phase III component

The FACT-G is a validated 27-item measure in which a higher score represents higher quality of life (QOL). Physical, functional, social and emotional well-being subscale scores are added together to form the FACT-G total score. Responses range from 0=Not a lot to 4=Very much. Certain items must be reversed before being added, by subtracting the response from 4. Subscale items are added together, multiplied by the number of items in the subscale, then divided by the number of items answered to obtain subscale totals. Total score ranges from 0-108; physical, social and functional subscales from 0-28; emotional subscale from 0-24. Each subscale requires at least 50% of the items to be completed while the overall response rate must be greater than 80%. If items are missing the subscale scores can be prorated. Change at 3 months is calculated as 3 month score - baseline score with a positive change indicating improvement in QOL.

Eligible patients randomized in the phase III component who consented to QOL having both baseline and 3 month FACT-G total score.

The Brief Pain Inventory Worst Pain score measures self-reported worst pain in the last week. Possible scores range from 0 (no pain) to 10 (worst pain imaginable), with higher scores indicating a worse outcome. Change is calculated as baseline score subtracted from post-baseline score with a positive change indicating more pain.

Questionnaires were not completed by all eligible participants randomized in the phase III component who consented to QOL. Therefore, the number of participants reported below are the number with the relevant questions answered at baseline and the given time point on each arm.

The EQ-5D index score measures health status. Possible scores range from 0 (worst health state) to 1 (best health state), with higher scores indicating better health. Change is calculated as baseline score subtracted from post-baseline score with a positive change indicating improved health status.

Questionnaires were not completed by all eligible participants randomized in the phase III component who consented to QOL.. Therefore, the number of participants reported below are the number with the relevant questions answered at baseline and the given time point on each arm.

The FACT-G total score gives a combined score of the four domains of health-related quality of life (HRQOL) in cancer patients: Physical, social, emotional, and functional well-being. Possible total scores range from 0 to 108 with higher scores indicating a better HRQOL. Change is calculated as baseline score subtracted from post-baseline score with a positive change indicating improvement in HRQOL.

Questionnaires were not completed by all eligible participants randomized in the phase III component who consented to QOL.. Therefore, the number of participants reported below are the number with the relevant questions answered at baseline and the given time point on each arm.